next | previous | table of contents

IX. Conclusion

This Report picks up where our June 2000 report left off: The death penalty in this country is a broken system that is of rising concern to many Americans. The public places great demands on the death penalty and yet has become increasingly aware that, as currently imposed, the penalty is a costly failure that does not serve the purposes for which it was established and risks taking the lives of innocent people. (See Part I above.)

Our earlier report documented these costs and risks. It showed that serious error is widespread and chronic. This is true no matter how conservatively one counts the number of judicially reversible mistakes the death penalty system makes. A review of our methods in this Report shows that we defined serious mistakes cautiously and counted them so conservatively that we excluded a number of death verdicts imposed on people who were innocent.970 Even defined this narrowly, capital error rates were 50% or more in nearly all death-sentencing states and years. Because such error keeps death verdicts from being carried out,971 this finding means that most states have failure rates above 50%. Nationally, the average failure rate is a nearly 70%; and capital verdicts in many states and counties fail at rates of 80%, 90% and even 100%. Over the 23-year course of our study from 1973 to 1995, barely 5% of the 5800 death verdicts that were imposed were carried out. During that period, the average time from death sentence to execution was 9 years. Today, given the exacting review needed to catch so much error, that delay averages 12 years. (See Parts III.A and B.)

Each one of the thousands of capital errors identified by state courts (which found 90% of the errors) and federal courts (which found the rest) is serious. This is true because each error stymies the execution of sentence at a cost of years of delay and hundreds of thousands or even millions of dollars in litigation costs. But it is more fundamentally true because reversible error is, by its very nature, serious error. Especially given the strong pressures on reviewing judges to approve even admittedly flawed verdicts, and given the strong bias of the rules governing court review towards approving verdicts, reversible error:

  • nearly always undermines the reliability of the verdict that the defendant committed a crime that was aggravated enough to warrant death as a punishment;

  • often risks the execution of people who are innocent of the crime or at least of the death penalty; and

  • always frustrates the demands and expectations of the public who adopted the death penalty, the taxpayers who pay for it and the victims who directly rely on it.

We have taken it as a research imperative, therefore, to identify the conditions and practices that are significantly linked to, and predict the occurrence of, serious capital error. (See Parts IIIC-E.)

The central object of this study is to discover information of use in answering two questions. Why is there so much error in capital cases? Can anything be done to solve the problem or at least to moderate the amount of serious error?

We use a two-part method for conducting this research. First, we design and carry out a single multiple regression analysis that makes the best use of our detailed data about factors that may predict where and when capital error rates are likely to be high. Then, we verify the reliability and results of this "best" analysis using a wide variety of alternative regression techniques, diagnostic tests for evaluating methods and results, categories of reversal rates being explained, and potentially explanatory factors operating at the state, county and case levels. (See Part IV.)

The bulk of this Report is a detailed presentation of the results of:

  • our main multiple regression analysis of explanations for the higher and lower rates of serious capital error in each of the 34 death-sentencing states that were active during the 23-year study period in each of the years in which they were active (see Parts VA, B, E);

  • seven follow-up regression analyses of those same state reversal rates—including ones examining only reversals at the state direct appeal, state post-conviction and federal habeas stages of review—to make sure our results reflect actual relationships in the data and are not products of particular research methods (see Parts VA-E);

  • 10 additional follow-up regression analyses of state-level and county-level explanations for different error rates in the 1002 active death-sentencing counties during each of the 23 years when they were active—including a study of counties in the three most active death-sentencing states during the study period (Florida, Georgia and Texas) (see Parts VIA-F);

  • case studies comparing rates of serious error, and rates of sentencing innocent defendants to die, in high and low death-sentencing counties (see Part VI.G.);

  • detailed case studies of four innocent individuals who were sentenced to die and whose capital verdicts were approved at all three review stages (see Part III.B.7.c ); and

  • a comprehensive case-level study of factors that predict reversals as opposed to affirmances of the 600 federal habeas verdicts that were fully reviewed during the period (see Part VII).

Based on these results we reach several overarching conclusions about conditions that predict the existence and high rates of serious capital error (see Part VIII):

  • Studying the problem of serious capital error using statistical and other techniques identifies a number of factors that predict high numbers or rates of capital reversals and are:

    • statistically significant, meaning there is only a small probability that they are the result of chance, as opposed to actual relationships between capital error and the identified explanations for it;

    • reliable in that they satisfy a number of diagnostic tests in most cases;

    • linked to sizeable differences in predicted rates of serious capital error, because, holding other factors constant at their averages, a fairly small change in a explanatory condition is associated with a fairly large increase or decrease in the amount or rates of serious error;

    • each part of a strong and coherent overarching explanation for serious capital error.

  • For the most part, the conditions our analyses link to sizeable differences in rates and amounts of serious capital error are capital-sentencing policies—how often, in response to what pressures, and in what broad classes or categories of cases, are death sentences sought and imposed—not traits of particular officials, jurors, lawyers, defendants or victims.

  • The principal conclusion of all of our analyses is that heavy use of the death penalty, especially when it sweeps in cases where the evidence supporting a capital verdict is not substantial, is a leading predictor of serious capital error. States and counties that use the death penalty more often per 1000 homicides are significantly more likely to have substantially higher rates of serious capital error than other jurisdictions. In particular, cases with low levels of aggravation that are swept into the capital category by jurisdictions' broad capital-sentencing policies and low capital-sentencing thresholds are prime candidates for serious, reversible error. Heavy use of the death penalty also leads to court congestion and delay in processing capital appeals.

  • Four other conditions strongly predict high rates of serious capital error. Each is either a measure of fears about serious crime, or a mechanism through which those fears can generate political pressure on officials to respond forcefully to crime, including through increased use of the death penalty. Some of those fears are based on actual crime and punishment rates. Others, more disturbingly, are sensitive to politics and race. We conclude that the tendency of all four conditions to heighten pressure to use the death penalty helps explain their link to high rates and amounts of serious capital error. The four conditions are:

    • the homicide threat to politically influential communities—measured by comparing the rates at which whites and blacks are victimized by homicides;

    • well-founded doubts about the ability of state law enforcement policies and officials to respond effectively to the problem of serious crime—measured by the rate at which serious criminals are apprehended, convicted and incarcerated;

    • state judges' susceptibility to negative political consequences if they do not conform their rulings in capital cases to popular sentiments—measured by the extent to which judicial selection techniques place state judges at risk of political discipline for unpopular rulings; and

    • the size of African-American and poor communities, which some influential citizens and officials evidently associate with higher rates of serious crime.

  • Underfunded and overburdened court systems—another consequence in part of high death-sentencing rates—also increase the risk of serious capital error.

  • Reviewing courts do not effectively keep serious errors from occurring or keep all unreliable death verdicts from being carried out.

    • The review process fails totally to prevent serious error from recurring.

    • It does not catch all, including some of the most serious, mistakes.

    • As a result, the probability that innocent people have been executed is high.

  • There is no reliable evidence that the conditions causing serious, reversible error have improved over time, and strong evidence that some of those conditions have gotten worse.

Having identified these death-sentencing policies that predict serious, reversible error in capital cases—and the political, economic and racial pressures that generate those policies—we next consider the reform options they suggest for addressing the chronically exorbitant amounts and rates of that error that have characterized the capital system for decades (Part VIII).

It is unlikely that policy changes can do more than moderate the problem of chronically high rates and amounts of serious capital error, the ill effects of error on the effective functioning of the death penalty system and the risk error creates of executing the innocent. This is because the same state and local policy makers who developed the aggressive death-sentencing thresholds and practices that so strongly predict serious error would have to be relied upon to adopt and maintain effectively ameliorative policies. And it is also because those policy makers will continue to face the same or growing fears about serious criminal behavior, and the same financial constraints and racially sensitive political pressures, that led them to adopt the risky policies in the first place.

In some states and counties, the costs and frustration levels associated with the death penalty may be so high that only a comprehensive solution to the problem of chronic capital error and its attendant costs and risks will suffice. In those places, the available options are to stop using the death penalty altogether, or to limit its use to a small number of offenses that are so highly aggravated that there is close to a social consensus that only the death penalty will serve.

For jurisdictions that prefer to explore more incremental solutions, at least in the short run, our study findings suggest 10 policy options for moderating serious capital error:

  • requiring proof beyond any doubt that the defendant committed the capital crime;

  • requiring that aggravating factors substantially outweigh mitigating ones before a death sentence may be imposed;

  • barring the death penalty for defendants with inherently extenuating conditions—mentally retarded persons, juveniles, severely mentally ill defendants;

  • making life imprisonment without parole an alternative to the death penalty and clearly informing juries of the option;

  • abolishing judge overrides of jury verdicts imposing life sentences;

  • using comparative review of murder sentences to identify what counts as "the worst of the worst" in the state, and overturning outlying death verdicts;

  • basing charging decisions in potentially capital cases on full and informed deliberations;

  • making all police and prosecution evidence bearing on guilt vs. innocence, and on aggravation vs. mitigation available to the jury at trial;

  • insulating capital-sentencing and appellate judges from political pressure; and

  • identifying, appointing and compensating capital defense counsel in ways that attract an adequate number of well-qualified lawyers to do the work.

Approaches that would likely magnify the amount of serious error are:

  • cutting back further on the scope of review of capital verdicts, which would likely increase the ill-effects of chronic error and invite more error;

  • making piecemeal additions to the list of qualifying aggravating circumstances;

  • shifting to the state the full costs of local capital prosecutions; and, most importantly

  • doing nothing.

The very last point is the most important one. Over decades and across dozens of states, large numbers and proportions of capital verdicts have been reversed because of serious error. The capital system is collapsing under the weight of that error, and the risk of executing the innocent is high. Now that explanations for the problem have been identified, and a range of options for responding to it are available, the time has come to fix the death penalty, or end it.

next | previous | table of contents